Scene: A floor plane rotated at 45deg. A cloth plane rotated at 45deg and 1m above the floor plane. The floor plane uses a Collision Modifier and the cloth plane uses a Cloth Modifier; when I run/play the simulation/animation the slanted cloth falls onto the slanted floor (collides), and slides down the slanted floor.

Question: How does Cloth Friction work in a scene such as this (or any other situation)? I tried increasing the Cloth Friction, from 0.0 to 80.0, and I didn't see any significant difference in the speed of the cloth sliding down the slanted floor. I did see a change in the speed when I changed the rotation on the cloth&floor from 45deg to 35deg and 25deg -- the cloth slides faster when the floor has a steeper incline. However, I didn't see a difference in speed/motion when I increased the Friction to its maximum (80.0). I'd love to try to increase the maximum past 80.0 to see if it works, but I don't think it is possible with Blender. Is this expected/normal behavior? I would expect the cloth to slide down slower with a higher friction.

  • $\begingroup$ The current version of Blender I'm using is 2.74 -- it seems like cloth is having some inconsistencies across earlier versions. Any ideas? $\endgroup$
    – Luis B
    Commented Jul 14, 2015 at 5:14

1 Answer 1


Here's a simulation of the situation (Blender 2.75).

enter image description here

The parameters that I cosidered in this "catalogue" are:

  • Resolution of surfaces (unsubdivided-> first row; Collision object subdivided-> second Row; Cloth object subdivided-> third row; Both subdivided-> fourth row)
  • Friction (0.1 or 80)
  • Repel Distance (0.001 or 1)

enter image description here

By comparing the elements of the first row, I can say that it seems to be at least a little difference of the behavior when friction jumps from 0.1 to 80 (but is not that strong as one expect)

Look at the last element of the matrix (4-4): it is the one where friction reveal itself in a more convincing way..but there's another, which has the same behavior: 2-4. This led me to think that is also important to consider how many faces your collision object has.

Now compare 2-4 with 2-2... also the max Repel distance influence the behavior.


It seems that changing only the Friction factor is not enough to run the sim as expected, but you have to deal at least with other two parameters:

  • resolution of collision object (should be enough detailed, but more subd won't give more frictional effect apart from the contribute given by the fact that the faces are in range for a longer period)
  • repulsion distance (act as a sort of detecting thershold, but more rep dist won't give more frictional effect)

Friction is the value that drives the pover of frictional effect, but it must be supported by other variables in order to affect the sim.

Here's the blend of where the gif comes from:

  • $\begingroup$ Thank you for putting the thought, clarity, and effort into this demonstration and explanation. This helps very much. I had subdivided my cloth plane, but not my floor plane. This makes sense, since it has more particles with which to calculate friction. $\endgroup$
    – Luis B
    Commented Jul 14, 2015 at 20:08
  • $\begingroup$ However, I do not understand how Repulsion Distance makes a difference for friction. How does higher repulsion distance result in higher friction? (I just assumed that the collision detection just happened at a greater "repulsion distance") $\endgroup$
    – Luis B
    Commented Jul 14, 2015 at 20:10
  • $\begingroup$ Saying that I'm not a coder but just a normal user (so I don't know what is the real math behind those parameters), I can say that, basing on my little experience, Rep Distange act as a far range..so if the cloth is much distant from the collision objet thant the value, friction is not considered anymore...I know that the name doesn't suggest this kind of behavior, this let me think that I am also missing something $\endgroup$
    – Carlo
    Commented Jul 14, 2015 at 20:21
  • $\begingroup$ Is not necessary to subdivide much, just wat you need to run a proper simulation. With the example discussed above, a collision object subdivided one time should be enough. $\endgroup$
    – Carlo
    Commented Jul 14, 2015 at 20:32
  • 1
    $\begingroup$ Very interesting: If I change the Repulsion Distance (aka, max distance) to 10.0, then the friction is higher; if I change it back to Repulsion Distance of 0.1, then the higher friction is maintained (same/similar to RD=10.0). It is as if increasing the range of the Repulsion Distance, allows the correctness of the friction/collision detection to increase. Is this the expected behavior? $\endgroup$
    – Luis B
    Commented Jul 14, 2015 at 23:40

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .